IMITATION/AI
Independent Agentic Research
Latest Paper: Philosophical Hyperparameters, Grounding LLMs via Configurable Ethical and Philosophical Frameworks
Today’s AI systems behave—but they don’t believe.
Their alignment is reactive, shaped by post-hoc constraints and surface-level safety tuning rather than stable, principled grounding.
Philosophical Hyperparameters introduces a new approach to AI alignment, treating ethical and philosophical frameworks not as abstract guidance, but as configurable control parameters for behaviour. Instead of training models toward a single, risk-averse norm, the paper demonstrates how large language models can be grounded in explicit ethical worldviews that act as a moral operating system.
The paper presents a proof-of-concept system in which LLMs are configured with 55 distinct philosophical frameworks, spanning classical ethics, religious traditions, and modern meta-ethical positions. Each configuration is evaluated across 20 complex dilemmas and quantitatively measured along 11 behavioural axes, including harm avoidance, risk appetite, autonomy, rule adherence, and collective good orientation. This produces a multidimensional behavioural signature for each framework, allowing alignment to be measured, compared, and audited rather than assumed.
The results show that LLMs can exhibit deep, coherent, and repeatable behavioural alignment when grounded in explicit philosophical convictions. These behaviours cluster into a small number of emergent moral regimes—such as harm-minimisation absolutism, risk-on activation, and rule-bound legalism—revealing a new design space for controllable, interpretable AI behaviour. The work reframes alignment as configuration rather than constraint, offering a path toward AI systems whose decisions are transparent, intentional, and anchored in known ethical postures rather than opaque optimisation pressures.
Cognitive Geometry for Agentic Memory Systems: Bridging Neural Correlation and Symbolic Reasoning
Today’s AI systems remember—but they don’t think.
Their memories function as static repositories, not dynamic structures of understanding.
Cognitive Geometry in Memory Systems introduces a new way to design memory for AI—one that transforms information into reasoning. The paper presents a four-layer cognitive memory model that maps different types of knowledge to unique geometries:
Facts → Knowledge Webs: semantic graphs that enable deep associative recall.
Experiences → Semantic Temporal Knowledge Webs: time-indexed graphs that preserve sequence, context, and meaning.
Procedures → Causal Trees: hierarchical chains of cause, effect, and subgoal.
Personalisation → Forests of Knowledge Trees: adaptive, individualised reasoning spaces.
This geometry creates a bridge between neural memory (pattern-based learning) and symbolic memory (logic-based reasoning)—allowing systems to store, transform, and reason over knowledge dynamically.
The result is a neuro-symbolic substrate that tackles long-standing LLM challenges: catastrophic forgetting, weak interpretability, and the inability to learn continuously. It reframes memory as the cognitive engine of agentic AI.
About Me:
-
My name is David Fearne, and I am an independent researcher and founder of Open Research at Imitation AI, a organisation dedicated to advancing the discussion around responsible design and deployment of artificial intelligence.
My research explores how AI can be introduced in ways that are beneficial to humanity, balancing technological capability with human values, ethics, and long-term societal responsibility. Rather than treating AI adoption as a series of disconnected experiments, my work focuses on building systematic approaches that organisations can use to integrate AI more effectively, sustainably, and with measurable alignment to their goals.
I am also Head of Research, Development and Innovation at NTT Data’s Global Data and AI practice, where my research is applied to advance real world applications for our clients.
-
Synthetic Reasoning
A framework for modelling human thought in a way that enables large language models to replicate chains of reasoning without falling into cognitive overload. This work explores how to break complex reasoning into controlled, focused heuristics, enabling AI to emulate both intuitive (System 1) and deliberative (System 2) forms of reasoning.The Agentic Enterprise
A model that applies the Viable System Model (VSM) to AI architecture, positioning AI agents as recursive organisational units. This research provides a blueprint for scaling agentic systems across enterprises while maintaining governance, adaptability, and strategic alignment.Graduated Memory Horizons & Graduated Intelligence
Theories that structure how AI agents should retain and apply knowledge across different levels of autonomy. By aligning memory and intelligence with organisational layers, these models support both immediate operational effectiveness and long-term organisational learning.
-
The ultimate aim of my research is to equip organisations with the intellectual tools and frameworks needed to embrace AI successfully—not as isolated “science experiments,” but as coherent, human-aligned systems. By distilling lessons from my own research, I work to ensure that AI is introduced in ways that are transparent, ethical, and deeply aligned with human progress.
Research and Opinions
Navigating the Path Between Transactional Myopia and Drowning in Infinite History
Applying Stafford Beer’s Viable System Model to create The Autonomous AI Organisations
The blending of Deterministic and Non Deterministic approaches in the Agentic Enterprise
Building the Agentic Enterprise
Synthetic Reasoning: A New Approach to Emulating Human Thought Processes in AI Systems
Unveiling the Concept of Cognitive Layers in Generative AI Applications
Invent the future today
It all begins with an idea join the our Open Research Forum
Info:
-
Contact our Open Research Forum.
info@imitation-ai.com
-
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).
To view a copy of this license, visit https://creativecommons.org/licenses/by-nc/4.0/
-
If you use this research in a publication, product, or derivative work, I kindly ask that you cite the original work and respect the spirit of open research.
This work is not to be used in military, surveillance, or exploitative commercial contexts.